2024 – Present
Sensory Feedback via Neuromorphic Tactile Sensing
Designed a piezoresistive tactile pipeline for robotic fingertips enabling real-time texture recognition across 21 surface conditions with low-latency, energy-efficient edge deployment. Led to the uSense paper under review to Nature Communications AI & Computing.
Live signal simulation
Neuromorphic
Tactile Sensing
Edge AI
Prosthetics
2024 – 2025
Robust Myoelectric Control via Latent Space Modeling
Designed and deployed a multimodal hand-gesture dataset via LSL synchronization. Developed a position-agnostic, day-agnostic, user-agnostic EMG decoder — published in IEEE TNSRE 2025.
sEMG signal stream
sEMG
BCI
Deep Learning
One-Shot Learning
2026 – Present
Multimodal Gesture Decoding: EMG + Ultrasound + IMU
Fusing ultrasound, sEMG, and IMU signals to build a generalizable gesture decoder. Investigating cross-modal correlations to relate peripheral muscle activity to cortical signals for next-gen prosthetic interfaces.
Accuracy by modality
Sensor Fusion
Ultrasound
IMU
Multimodal ML
2021 – 2022
EEG Emotion Recognition via Riemannian Geometry
Increased emotion recognition accuracy by 9% using lower-dimensional Riemannian geometry. Improved feature selection achieving +5% accuracy gain. Published in Applied Sciences 2023.
Feature cluster map
EEG
Signal Processing
Riemannian Geometry
Emotion AI
2023
Autonomous Patient Transport Robot
Implemented Visual SLAM with Intel RealSense for safe indoor navigation. Led system-level design across sensing, perception, and evaluation. Mentored undergraduates in ROS and Linux development.
Sensor signal trace
ROS
SLAM
HRI
Robotics
2019 – 2021
Agricultural Robot: Disease Detection & Navigation
Built CNN-based crop disease and ripeness classification. Achieved collision-free navigation using HRI. Annotated tomato ripeness dataset; presented at IEEE ICCAS & IEMCON.
Classification accuracy
Computer Vision
CNN
ROS
Autonomous Navigation
